The use of Toll-like receptor ligands as adjuvants in fish vaccines

(1)

as adjuvants in fish vaccines

Thesis submitted for the degree of Philosophiae Doctor by

Marianne Arnemo

Department of Pharmaceutical Biosciences School of Pharmacy

Faculty of Mathematics and Natural Sciences University of Oslo

201 6

(2)

Series of dissertations submitted to the

Faculty of Mathematics and Natural Sciences, University of Oslo No. 1725

ISSN 1501-7710

reproduced or transmitted, in any form or by any means, without permission.

Cover: Hanne Baadsgaard Utigard.

Print production: Reprosentralen, University of Oslo.

(3)

The work presented in this thesis was carried out at the Department of Pharmaceutical Biosciences, School of Pharmacy at the University of Oslo from August 2011 to December 2015 under the supervision of Professor Tor Gjøen. Associate professor Hanne C. Winther- Larsen served the role as co-supervisor.

First and foremost I want to thank my main supervisor Tor Gjøen for giving me the opportunity to participate in the project. Thank you for your encouragement and skilful guidance throughout the PhD. Thank you for being an amazing supervisor.

Special thanks to my research family (the “zebrafish family”): Tor Gjøen, Anne-Lise Rishovd, Arturas Kavaliauskis, and Adriana Magalhaes Santos Andresen. It has been a pleasure working with you. Thank you for also being my friends and for all support, help, humour, and kindness. You have been very important for me during the PhD and I will miss having you as a part of my daily life.

I thank Alexander Rebl, Sonia Dios, Beatriz Novoa, Bente Ruyter, Marta Bou Mira, Øystein Evensen, and the rest of the co-authors for valuable and fruitful collaboration projects.

I thank all my ZEB-colleagues that contributed to a nice work environment in the office and in the laboratory.

I thank my Dad Jon Martin Arnemo and Veronica Sahlén for help proof-reading my thesis.

I am grateful to my closest family (Mum, Dad, Cecilie, and Nils Olav) and all my friends for support, love, and help through good and difficult times. Thank you all.

Thank you Emil, for being here for me with your support, patience, and love.

Thank you Ole, for everything - I will miss you forever.

Marianne Arnemo Blindern, December, 2015

(4)

(5)

5

List of publications

Paper I:

Effects of TLR agonists and viral infection on cytokine and TLR expression in Atlantic salmon (Salmo salar).

Arnemo M, Kavaliauskis A, Gjøen T. Developmental and Comparative Immunology (2014). 46, 139-145. doi:10.1016/j.dci.2014.03.023

Paper II:

Structurally diverse genes encode Tlr2 in rainbow trout: The conserved receptor cannot be stimulated by classical ligands to activate NF-kappaB in vitro.

Brietzke A, Arnemo M, Gjøen T, Rebl H, Korytář T, Goldammer T, Rebl A, Seyfert HM.

Developmental and Comparative Immunology (2016). 54, 75-88.

doi:10.1016/j.dci.2015.08.012 Paper III:

Use of poly I:C stabilised with chitosan as a vaccine-adjuvant against Viral Haemorrhagic Septicaemia Virus infection in zebrafish.

Kavaliauskis A, Arnemo M, Kim SH, Ulanova L, Speth M, Novoa B, Dios S, Evensen Ø, Griffiths GW, Gjøen T. Zebrafish (2015). 12, 421-431. doi:10.1089/zeb.2015.1126 Paper IV:

Chitosan-poly I:C nanoparticles: a novel adjuvant in aquaculture vaccines. A study of particle bio- distribution and immune response in zebrafish (Danio rerio).

Kavaliauskis A, Arnemo M, Speth MT, Lagos Rojas LX, Rishovd AL, Estepa A, Griffiths G, Gjøen T. Manuscript submitted for publication.

Paper V:

Effects of dietary n-3 fatty acids on Toll-like receptor activation in primary leucocytes from Atlantic salmon (Salmo salar).

Arnemo M, Kavaliauskis A, Mira MB, Berge GM, Ruyter B, Gjøen T. Manuscript submitted for publication.

(8)

8

Additional scientific work from the PhD period (not included in this thesis)

Preclinical immunogenicity and functional activity studies of an A+W meningococcal outer membrane vesicle (OMV) vaccine and comparisons with existing

meningococcal conjugate- and polysaccharide vaccines.

Tunheim G, Arnemo M, Næss LM, Fjeldheim ÅK, Nome L, Bolstad K, Aase A,

Mandiarote A, González H, González D, García L, Cardoso D, Norheim G, Rosenqvist E.

Vaccine (2013), 31, 6097-106. doi:10.1016/j.vaccine.2013.09.044

Activation of unfolded protein response pathway during infectious salmon anemia virus (ISAV) infection in vitro an in vivo.

Kavaliauskis A, Arnemo M, Rishovd AL, Gjøen T. Developmental and Comparative Immunology (2016). 54, 46–54. doi:10.1016/j.dci.2015.08.009

Stability of a Vesicular Stomatitis Virus–Vectored Ebola Vaccine.

Arnemo M, Viksmoen Watle SS, Schoultz KM, Vainio K, Norheim G, Moorthy V, Fast P, Røttingen JA, Gjøen T. Journal of Infectious Diseases (2015). Published online November 12,2015. doi:10.1093/infdis/jiv532

Effects of doses, aluminium hydroxide as adjuvant and mouse strain on immune responses in mice immunised with a meningococcal A+W outer membrane vesicle (OMV) vaccine.

Tunheim G, Arnemo M, Bolstad K, Sinnadurai K, Fjeldheim ÅK, Næss LM, Norheim G, Mandiarote A, Gonzalez D, Garcia L, Gjøen T, Rosenqvist E. Manuscript submitted for publication.

(9)

9

Summary of thesis

Non-living antigens are often poorly immunogenic and require addition of adjuvants to elicit protective immunity. Due to the immunostimulatory potential of Toll-like receptor (TLR) ligands, they are explored as vaccine adjuvants. The development of efficient and cheap vaccines against aquatic viruses is important for a sustainable aquaculture industry and the adjuvants for fish vaccines need to be improved. However, increased knowledge of fish TLR function is required before their ligands can find their way into fish vaccines.

The major aim if this thesis has been to contribute to a more detailed understanding of fish TLRs. First, the tissue distribution of all known Atlantic salmon TLRs, the immunostimulatory potential of a panel of TLR ligands in primary head kidney leucocytes, and the impact of viral infection on TLR expression in head kidney were investigated.

Head kidney and spleen were the main TLR expressing organs in Atlantic salmon. Several TLR ligands induced expression of inflammatory cytokines in salmon head kidney leucocytes. TLR3, TLR7, and TLR8a1 were induced in vivo after viral infection. In order to functionally validate ligand-specific activation of fish TLRs, we established an in vitro reporter assay in a salmon cell line. However, classical TLR2 ligands failed to activate rainbow trout TLR2 signalling when using NF-κB activation as measure of activation. To test the in vivo immunostimulatory potential of a TLR ligand alone and in vaccine formulations, a cold-water zebrafish challenge model was used. The TLR3 ligand poly I:C induced expression of antiviral transcripts in zebrafish head kidney and pre-treatment with poly I:C delayed VHSV (viral haemorrhagic septicaemia virus)-induced mortality.

Chitosan encapsulated poly I:C was demonstrated to provide protection against VHSV when co-injected with two different non-living antigens (inactivated whole VHSV and VHSV glycoprotein G). Due to decreasing levels of the dietary n-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) in Atlantic salmon feed, we investigated how minimal levels of these fatty acids affect TLR signalling in Atlantic salmon leucocytes. The ability of leucocytes to respond to TLR ligand stimuli was reduced with low dietary- and head kidney levels of EPA and DHA, indicating the importance of n- 3 fatty acids in resistance to infection and response to vaccines.

Our results provide new knowledge in the fish TLR field and lend support to poly I:C as a promising adjuvant candidate in viral vaccines.

(10)

10

Sammendrag (Summary in Norwegian)

Adjuvanser er en gruppe substanser som ofte brukes for å øke immunogenisiteten til vaksineantigener uten egen replikasjon (inaktiverte virus og rekombinante proteiner).

Ligander som gjenkjennes av Toll-lignende reseptorer (TLR) stimulerer immunstemet og utforskes som potensielle nye adjuvanser i humane vaksiner. Det er behov for bedre virusvaksiner og nye adjuvanser til bruk i den globale akvakulturnæringen. For å kunne fullt utnytte den stadig økende informasjon om TLR og TLR-ligander i vaksiner til andre arter, må man øke kunnskapen om TLR i fisk. Hovedmålet med denne avhandlingen var derfor å bidra til økt forståelse om TLR i fisk. I artikkel 1, undersøkte vi distribusjonen av alle kjente TLR i atlantisk laks, hvordan hodenyre leukocytter responderte på stimulering med TLR-ligander og hvordan genuttrykket av TLR ble endret under en virusinfeksjon.

Hodenyre og milt var de organene som uttrykte høyeste nivåer av de fleste TLR, og flere av TLR-ligandene induserte økt uttrykk av inflammatoriske cytokiner i leukocytter fra hodenyre. TLR3, TLR7 og TLR8a1 ble oppregulert ved virussykdommen infeksiøs lakseanemi i laks. For å kunne måle ligandbinding til TLR fra fisk (artikkel 2), etablert et cellesystem basert på målinger av NF-κB-aktivitet (et nedstrøms signalprotein). TLR2 fra regnbueørret lot seg ikke aktivere med klassiske TLR2-ligander i dette systemet. Sebrafisk ble også brukt for å teste immunstimulerende effekt av en TLR-ligand og til utprøving av vaksiner basert på TLR-ligand formulert i nanopartikler (artikkel 3 og 4). TLR3-liganden poly I:C økte uttrykket av flere immungener i sebrafisk som er viktige i bekjempelsen av virussykdommer, samt at løselig poly I:C forsinket dødeligheten etter at fiskene var infisert med VHS (viral hemoragisk septikemi)-virus. Vaksiner som inneholdt poly I:C innkapslet i chitosan partikler kombinert med enten et inaktivert VHS-virus eller glykoprotein fra VHS-virus beskyttet sebrafisken fra VHS. Dette viser at poly I:C er en lovende adjuvans i virusvaksiner. På grunn av økt etterspursel etter eicosapentaensyre (EPA) og docosahexaensyre (DHA), har innholdet av disse omega-3 fettsyrene i laksefôr blitt kraftig redusert. Vi undersøkte derfor hvordan minimale nivåer av disse fettsyrene påvirker leukocytter fra atlantisk laks. Leukocytter fra gruppen som ble fôret med de laveste nivåene av EPA og DHA viste redusert evne til å respondere på TLR-ligander. Dette indikerer viktigheten av omega-3 fettsyrer for å bekjempe infeksjoner og evnen til å respondere optimalt på vaksiner.

(11)

11

Abbreviations

AP1 Activator protein 1 CLR C-type lectin-like receptors

COX Cyclooxygenase

CREB Cyclic AMP-responsive element-binding protein DHA Docosahexaenoic acid

dsRNA double stranded RNA EPA Eicosapentaenoic acid FCA Freund’s complete adjuvants FIA Freund’s incomplete adjuvants GALT Gut-associated lymphoid tissue GSK Glaxo Smith Kline

IFIT Interferon-induced proteins with tetratricopeptide repeats

IFN Interferon

IL Interleukin

ILT Interbranchial lymphoid tissue IPN infectious pancreatic necrosis IPNV Infectious pancreatic necrosis virus IRAK IL-1R-associated kinase

IRF Interferon regulatory factor ISA infectious salmon anaemia ISAV Infectious salmon anaemia virus ISG IFN-stimulated gene LBP LPS-binding protein

LGP2 Laboratory of genetics and physiology 2 LPS Lipopolysaccharide

LRR leucine-rich region LTA Lipoteichoic acid LTB4 Leukotriene B4 MAL MyD88-adaptor-like

MAPK Mitogen-activated protein kinases

MDA5 Melanoma differentiation-associated gene 5

(12)

12

MHC major histocompatibility complex MPL 3-O-desacyl-4'-monofosforyl lipid A

Myd88 Myeloid differentiation primary response gene (88) NF-κB Nuclear factor kappa B

NLR Nucleotide oligomerization domain like receptors OIE The world organization for animal health ODN Oligodeoxynucleotides

PAMP Pathogen associated molecular patterns PD Pancreas disease

PGE2 Prostaglandin E2 PGN Peptide glycan

PGRP Peptide glycan recognition protein PLA2 Phospholipase A2

PLGA Poly-(lactide-co-glycolide) Poly I:C Polyinosine–polycytidylic acid PRR Pathogen recognition receptor RIG-1 Retinoic acid-inducible gene-1

RLR Retinoic acid-inducible gene-1 like receptors SARM Sterile α-and armadillo-motif-containing protein SAV Salmon alphavirus

ssRNA Single stranded RNA Th1 Type 1 T helper cell Th17 Type 17 T helper cell Th2 Type 2 T helper cell TIR Toll/interleukin-1 receptor TLR Toll-like receptor TNF Tumor necrosis factor

TRAF TNF-receptor-associated factors TRAM TRIF-related adaptor molecule

TRIF TIR-domain-containing adaptor protein inducing IFN-β VHS Viral haemorrhagic septicaemia

VHSV Viral haemorrhagic septicaemia virus VLP Virus like particle

(13)

13

Introduction

1. Vaccines

1.1. A brief history of vaccines

During the last century, vaccination has had a tremendous impact on global health by reducing death and morbidity caused by infectious diseases. Vaccines are biological preparations that enhance immunity against disease and either prevent (prophylactic vaccines) or treat disease (therapeutic vaccines) (Delany et al., 2014). The principle of vaccination was first applied over 1000 years ago via the process of “variolation”; the inoculation of pustules from smallpox patients into healthy individuals who were then subsequently protected against the disease (Riedel, 2005). The British vaccine pioneer Edward Jenner developed the process further in the 1790s by showing that exposure to cowpox induces protective immunity to smallpox (the term “vaccination” is derived from the Latin words for cow and cowpox - vacca and vaccinia). This discovery led to a decline in smallpox mortality and, many years later, the eradication of smallpox in 1977 (Riedel, 2005, Minor, 2015). Since then there have been major advances in vaccine development.

Louis Pasteur’s principles “isolate, inactivate and inject” in the late 1800s led to the development of successful vaccines, based on inactivated toxins and live attenuated or inactivated (killed) pathogens, against many serious infectious diseases (Rappuoli, 2007, Plotkin, 2005). From 1950, many new effective inactivated, live attenuated, and subunits vaccines have been developed as a result of the progress in microbiology and gene technology.

1.2. Human viral vaccines

Live attenuated vaccines against viral diseases are one of the most cost effective health interventions currently available (Bloom et al., 2005). Poliomyelitis, measles, mumps, yellow fever, and rubella are examples of diseases that can be prevented by licensed live attenuated vaccines (Minor, 2015). Poliomyelitis is nearly eradicated, and measles and mumps are controlled in the western world (Minor, 2015). Live vaccines are generally very effective and induce long-lived immunity after only one single dose. Attenuation can be achieved by serial passages of the virus in cultured cells, applying harsh condition on a

(14)

14

virus strain, or using recombinant gene technology (Rappuoli et al., 2009, Kallerup and Foged, 2015). The viral strain will accumulate mutations that make it non-pathogenic, while it still possesses patterns of the original virus and mimics the natural infection by inducing an immune response (Clem, 2011). Today, the further development of these vaccines is limited by several safety concerns, e.g. risk of reversion to the virulent strain, disease in immunocompromised individuals, and spread of the attenuated virus into the environment (Lauring et al., 2010).

The safety concerns of live attenuated vaccines have led to a shift towards inactivated viruses or viral subunits as vaccines. Inactivated vaccines are generally less immunogenic than their live attenuated counterparts due to the lack of replication and fast clearance from the injection site; hence the often need for an additional booster dose and adjuvants.

However, such vaccines are more stable and safer than live vaccines. Inactivated vaccines are produced by viral cultivation to produce large quantities of the antigen and then inactivation by radiation, heat, or chemical agents. Inactivated vaccines usually do not require refrigeration, and they can be easily stored and transported in freeze-dried form, which makes them more accessible to people in developing countries (Sanders et al., 2015).

Subunit vaccines contain one or more components of a pathogen rather than the entire pathogen (like a protein, polysaccharide, glycoprotein, inactivated toxin, or outer membrane vesicle). The antigens are chosen because of their ability to elicit protective immunity. Production is more easily controlled and they offer considerable advantages over the inactivated and attenuated vaccines in terms of safety. Because of the lack of many pathogen features, these subunit vaccines are weakly immunogenic and often require co-administration of an adjuvant (Kallerup and Foged, 2015). Virus-like particles (VLPs) are derived from self-assembling subunits of viral structural proteins that mimic the structure of an authentic virus, but lack the viral genome. Vaccines consisting of VLPs combine the advantages of whole virus vaccines (strong immune response) and recombinant subunit vaccines (safe and simple vaccine). Several licensed VLP vaccines are available, e.g. against human papilloma virus and hepatitis B virus. These vaccines consist of one or more viral proteins (expressed in yeast, insect, or mammalian cells) that are important for inducing antibodies against the virus (Roldão et al., 2010).

(15)

15

DNA vaccines deliver genetic material that encodes a specific antigen to the skin or the muscle. The DNA enters local cells and uses the host cellular machinery to express the antigen encoded in the DNA vaccine. The synthesised antigen can then be presented in the major histocompatibility complex (MHC) for T and B cells to initiate immune responses.

Two viral DNA vaccines are licensed (vaccines against West Nile virus and infectious hematopoietic necrosis virus for use in horses and salmon respectively), and several DNA vaccines are currently tested in clinical trials (Kutzler and Weiner, 2008). It is also possible to use replicating or non-replicating viruses as vaccine vectors. The virus vector is a non- pathogenic virus that has been modified to encode and present antigens from a pathogen. A wide range of innovative viral vectors that are able to deliver antigens and induce immune responses are available (e.g. pox viruses, adenovirus, coronavirus, flavivirus, influenza virus, rhabdovirus etc. (Draper and Heeney, 2010).

1.3. The need for viral vaccines in fish farming

Aquaculture (farming of aquatic organisms) is a rapidly growing global industry and an important nutritional and economical source for many countries around the world, especially in Asia and South America. The aquaculture industry is also of great importance to Norway. In the late 1960s, salmon farming started in Norway to support rural fishing communities that were having economic problems due to reduction in wild-capture fishing activity (Liu et al., 2011). Since then the aquaculture industry has overcome many technical and biological challenges and has become one of Norway’s biggest export trades after oil and gas. The main species of Norwegian aquaculture today are salmonid fish (Atlantic salmon Salmo salar and rainbow trout Oncorhynchus mykiss). Today, Norway is the largest producer and exporter of Atlantic salmon globally, followed by Chile, United Kingdom, and Canada. In 2014, the total amount of produced Atlantic salmon in Norway exceeded 1.2 million tonnes, which constituted over 50 % of the total world production of this fish (Guttormsen, 2015).

Although aquaculture is one of the fastest growing food-producing industries in the world, there are still challenges that pose a threat to a sustainable growth of this industry. One of the major challenges is infectious diseases caused by bacteria, viruses, and parasites, whose detrimental impacts are facilitated by the effectiveness of pathogen transportation in water and the high density of fish in large-scale farming. Diseases can lead to great production losses, unacceptable animal welfare situations, and spread of disease to wild

(16)

16

fish in the area. Worldwide the most common pathogens causing infectious diseases are bacteria (over 50%), followed by viruses, parasites and fungi (Kibenge et al., 2012).

Today, in Northern Europe and North America, bacterial diseases are controlled by vaccines in most salmonid farms and the use of antibiotics is limited (Sommerset et al., 2005). However, viral diseases have been more difficult to control due to lack of antiviral therapeutics, high susceptibility of fish during the early stages of the life cycle, and insufficient knowledge about pathogenesis and natural resistance to viral infections. The development of efficient viral vaccine has also been a challenge (Kibenge et al., 2012).

Both farmed and wild fish are susceptible to a long list of viral pathogens. Some of the most important viral diseases and the causative viruses affecting farmed fish are listed in Table 1 (Kibenge et al., 2012, Crane and Hyatt, 2011, Dhar et al., 2014, Shoemaker et al., 2015). Eight viral fish diseases are listed as reportable diseases by the World Organization for Animal Health (OIE) in 2015 due to the risk of viral spread through commercial trade of fish and fish products (see Table 1) (Dhar et al., 2014). In Norway, pancreas disease (PD), heart and skeletal muscle inflammation (HSMB), infectious pancreatic necrosis (IPN), cardiomyopathy syndrome (CMS), and infectious salmon anaemia (ISA) are the most frequent viral diseases detected in farmed fish (Veterinærinstituttet, 2015).

Viral haemorrhagic septicaemia (VHS) is one of the oldest prevailing and most economically important viral diseases of salmonid fish in Europe and flounder in Asia. No vaccine is available, and the disease is reportable to OIE indicating the importance of this virus. The causative agent of VHS, viral haemorrhagic septicaemia virus (VHSV) is a single stranded RNA virus and member of the family Rhabdoviridae and genus Novirhabdovirus. The virus has five major structural proteins (nucleocapsid-, phospho-, matrix-, glyco- and RNA polymerase protein) and there are currently four genotypes identified (genotype I-IV) (Einer-Jensen et al., 2004, Skall et al., 2005). The most susceptible fish species is the rainbow trout, but the virus has since the first identification in the early 1900s been isolated from numerous wild and farmed fish species (Skall et al., 2005). VHSV causes high mortality rates (up to 100% in fry) and huge economical losses (Crane and Hyatt, 2011, Micol et al., 2005).

(17)

17

1.4. Current status of fish vaccines

The success of bacterial vaccines in fish have led to a decrease in the use of antibiotics, and several vaccines against bacterial diseases are used in aquaculture worldwide (Håstein et al., 2005). In Norway, bacterial diseases caused enormous losses to the salmon farming during the 1980s and tonnes of antibiotics were used (Figure 1). However, due to the introduction of effective bacterial vaccines and improved health management, the total consumption of antimicrobials was reduced by 99% and made it possible for the huge increase in production (NORM/NORM-VET2013, 2013). Today in Norway, the salmonid population is vaccinated against three major bacterial diseases (vibriosis, cold-water vibriosis, and furunculosis) before release into sea water. Worldwide vaccination has been most important in salmonids and species like sea bass (Dicentrarchus labrax) and sea bream (Sparus aurata) (Håstein et al., 2005). The main administration routes for fish are injection and immersion. Oral vaccines have also been used, but have not provided the same efficiency as immersion and injectable vaccines. The bacterial vaccines in use are simple, consisting of formalin-inactivated whole bacteria. Many of the registered vaccines Table 1: List of viral diseases impacting farmed fish

Disease Causative virus Virus family Host examples

Infectious pancreatic necrosis

(IPN)* Infectious pancreatic necrosis virus

(IPNV) Birnaviridae Salmonids, halibut,

common carp Viral encephalopathy and

retinopathy (VER) or Viral nervous necrosis (VNN)

Nervous necrosis virus (NNV) Nodaviridae Atlantic halibut, Atlantic cod, sea bass, grouper Infectious salmon anaemia

(ISA)*^a Infectious salmon anaemia virus

(IAV) Orthomyxoviridae Salmonids

Pancreas disease (PD) or

sleeping disease (SD)*^a Salmon alpha virus (SAV) Togaviridae Atlantic salmon, rainbow trout Infectious hematopoietic

necrosis (IHN)^a Infectious hematopoietic necrosis

virus (IHNV) Rhabdoviridae Salmonids,

sturgeon, herring Epizootic hematopoietic

necrosis (EHN)^a Epizootic hematopoietic necrosis

virus (EHNV) Iridoviridae Rainbow trout,

perch species Viral haemorrhagic

septicaemia (VHS)*^a

Viral haemorrhagic septicaemia virus (VHSV)

Rhabdoviridae Rainbow trout, turbot, flounder Spring viremia of carp (SVC)^a Spring viremia of carp virus (SVCV) Rhabdoviridae Carp species Cardiomyopathy syndrome

(CMS)* Piscine myocarditis virus (PMCV) Totiviridae Atlantic salmon Heart and skeletal muscle

inflammation (HSMI)*

Piscine reovirus (PRV) (suspected) Reoviridae Atlantic salmon Koi herpesvirus disease

(KHVD)^a Koi herpesvirus (KHV) Alloherpesviridae Common carp, Koi Red sea bream iridoviral

disease (RSID)^a Red sea bream iridovirus (RSIV) Iridoviridae Sea bream species

*Diseases reported in Norway.

a Listed as reportable fish diseases by The World Organisation for Animal Health (OIE).

References (Kibenge et al., 2012, Crane and Hyatt, 2011, Dhar et al., 2014, Shoemaker et al., 2015)

(18)

18

are multivalent, i.e. they protect fish against more than one bacterial disease. To obtain satisfactory protection, adjuvants are included in the vaccines (Håstein et al., 2005).

Figure 1: Amount of antibiotics (tonnes) for therapeutic use in farmed fish in Norway versus produced biomass farmed fish (1,000 tonnes) (NORM/NORM-VET2013, 2013). The arrows mark the introduction of bacterial vaccines (Sommerset et al., 2005).

Although vaccination against diseases in aquaculture has enabled control of many bacterial diseases, the development of efficient and cheap vaccines against viral diseases has proven very challenging. The few existing vaccines for viral diseases in fish are either monovalent vaccines or included in multivalent bacterial vaccines. Live attenuated vaccines induce strong and sustained immune responses to the target disease in fish, but there are environmental and regulatory concerns hampering their further development. Most aquaculture operations are without physical barriers to wild-living fish in the same area and strains attenuated for aquaculture species may be virulent in wild species living in farm areas (Dhar et al., 2014, Salgado-Miranda et al., 2013, Brudeseth et al., 2013).

The majority of the commercial viral vaccines are based on inactivated virus and target salmonid viruses like infectious pancreatic necrosis virus (IPNV), infectious salmon anaemia virus (ISAV), and salmon alphavirus (SAV) (Dhar et al., 2014). However, many

(19)

19

attempts to make inactivated viral vaccines (although adjuvanted) have failed due to low immunogenicity. In addition, some viruses display poor or non-existent replication in cell cultures, making it difficult to grow sufficient quantities of antigen for vaccine production.

A few subunit vaccines are available against IPN and ISA, and consist of a capsid protein (VP2) of IPNV and the ISAV recombinant hemagglutinin esterase gene, respectively. One DNA vaccine, based on the gene of IHNV glycoprotein, is licensed in Canada, but is not approved in Europe or the United States due to safety concerns regarding the integration of foreign genes in food (Evensen and Leong, 2013). Although the licensed viral vaccines have been documented to be effective in experimental trials, the efficacy of these vaccines under field condition is uncertain due to a lack of published reports and continued occurrence of viral outbreaks (Rimstad, 2014, Robertsen, 2011, Kibenge et al., 2012, Rimstad, 2011, Veterinærinstituttet, 2015).

Like the inactivated bacterial vaccines, the inactivated virus and viral subunits are non- replicating antigens resulting in lower immunogenicity and often need to be accompanied by adjuvants. The oil-based adjuvants included in the available fish vaccines can lead to severe adverse effects (covered in section 2.2.). The list of subunit- (both recombinant protein and VLP) and DNA vaccines under development for fish is long, but the success of these vaccines is dependent on adjuvant improvements (Tafalla et al., 2013).

2. Vaccine adjuvants

2.1. Development of adjuvants for use in humans

Non-living vaccine antigens (inactivated- and subunit vaccines) are often poorly immunogenic and will require development of adjuvants that can stimulate induction of protective humoral- and cell-mediated immunity (Coffman et al., 2010). Adjuvants (from latin adjuvare meaning “to help”) are a class of substances with a shared feature of increasing the immunogenicity and the efficacy of vaccines. The principle was first used by Gaston Ramon at Institute Pasteur in the 1920s and traditional adjuvants have been developed empirically, without a clear understanding of cellular and molecular mechanisms of action (Kenney and Cross, 2009). Vaccine adjuvants are a heterogeneous group of substances with a wide variety of mechanisms of action and recent research on vaccine development has focused on adjuvant improvement. Recent evidence suggests that adjuvants work through one or more of the following mechanisms: sustained release of

(20)

20

antigen, upregulation of cytokines and chemokines, increased antigen uptake and presentation, activation of inflammasomes, and activation, maturation and migration of antigen presenting cells (Cox and Coulter, 1997, Awate et al., 2013).

The dominating adjuvants in licensed vaccines are aluminium salts (aluminium hydroxide or aluminium phosphate). Although aluminium salts have been used in vaccines for over 80 years, their mechanism of action is not completely understood. Depot effect (sustained release) and inflammasome activation are among the proposed mechanisms (Kool et al., 2008, Marrack et al., 2009). However, it is well-known that aluminium adjuvants induce production of local cytokines and chemokines, increase cell recruitment and antigen presentation that induce a type 2 T helper cell (Th2) skewed response, and increase antibody production (Marrack et al., 2009, Awate et al., 2013). Other adjuvants are based on oils (often emulsions), liposomes, microparticles, surface-active agents, pathogen- and plant derivatives, vitamins, or cytokines. However, most of them yet to be included in licensed vaccines and remain in experimental use (Kenney and Cross, 2009).

The benefits of using an adjuvant are many, e.g. increased antibody titre and protective immunity, dose sparing, reduced number of immunisations (increased immunological memory), increased effect in populations with low response (e.g. elderly and children), enabling a more rapid immune response (post-exposure prophylaxis), antibody response broadening, and induction of a desired immune response against the specific pathogen (e.g.

cell-mediated versus humoral response) (Coffman et al., 2010, Reed et al., 2013).

A shift from empiricism to rational design of adjuvants in human vaccinology research has led to several new, efficient adjuvants in vaccines that are already licensed or currently in clinical testing. A common feature of many new adjuvants under development is that they stimulate pattern recognition receptors (PRRs) expressed by innate immune cells. Various families of PRRs are identified, including Toll-like receptors (TLRs), C-type lectin-like receptors (CLRs), nucleotide oligomerization domain (NOD) like receptors (NLRs), and retinoic acid-inducible gene-1 (RIG-1) like receptors (RLRs) (Awate et al., 2013). These receptors sense conserved microbial features collectively called pathogen associated molecular patterns (PAMPs), and initiate innate immune responses as well as set the stage for an efficient adaptive immune response (Medzhitov, 2007). The best characterised family of PRR is the TLR family and their ligands’ ability to induce inflammatory cytokines is explored for the immunostimulatory and adjuvant potential (Awate et al.,

(21)

21

2013, Steinhagen et al., 2011, Coffman et al., 2010). 3-O-desacyl-4'-monofosforyl lipid A (MPL) is a detoxified derivative of lipopolysaccharide (LPS) and ligand for TLR4. MPL is part of Glaxo Smith Kline’s (GSK) Adjuvant System 04 (AS04) and is currently used in licensed vaccines against human papilloma virus and hepatitis B (Garçon et al., 2007).

TLR ligands as potential vaccine adjuvants are covered in section 4.4.

2.2. Adjuvants for fish vaccines

The most common adjuvants in fish vaccines are based on mineral oil emulsions, which have been more or less unchanged since the development of these vaccines. These adjuvants increase immunogenicity of the antigen by making a depot at the injection site from which the antigen is slowly released (Tafalla et al., 2013). The emulsion usually consists of a water phase with the antigen dispersed in an oil phase (usually a mineral oil) with a surfactant (e.g. mannitol oleate) to stabilise the emulsion. The adjuvants registered under the trademark Montanide are based on mineral oil, non-mineral oil, or a mixture of both, and have been used in several commercialised fish vaccines. Freund’s complete (FCA) and incomplete (FIA) adjuvants (known from human adjuvant research) consist of surfactant and paraffin oil with or without heat-killed mycobacteria. Both have been tested in experimental fish vaccines with variable results, but have not yet been used in commercial vaccines (Tafalla et al., 2013).

The side effects of oil-adjuvanted injection vaccines are undesirable for animal welfare reasons. Injectable vaccines formulated with oil-adjuvants can cause tissue lesions with granulomas at the injection site and abdominal cavity, adhesions between internal organs, autoimmunity reactions, reduced appetite and growth, and malfunction of reproductive organs (Koppang et al., 2004, Poppe and Breck, 1997, Håstein et al., 2005, Koppang et al., 2005, Haugarvoll et al., 2010, Midtlyng and Lillehaug, 1998). In addition to the type of adjuvants used, water temperature, fish size, hygiene during handling, time of the year, and anaesthesia can also influence the development and severity of side-effects (Berg et al., 2006, Håstein et al., 2005).

Although lagging behind the human adjuvant research, there are also promising adjuvant candidates for fish vaccines currently under development. For example, biocompatible and biodegradable nano- and microparticles offer a promising alternative to oil emulsions. The antigen can be covalently bound to or physically entrapped in these particles. Formulations based on polymers like poly-(lactide-co-glycolide) (PLGA) and chitosan are tested as

(22)

22

adjuvant systems in several fish species, both in injectable and oral vaccines (Plant and LaPatra, 2011, Tafalla et al., 2013). Trough cloning studies and fish genome projects, increasing knowledge about the fish innate immune system and pathogen recognition (e.g.

TLRs) have made it possible to move to adjuvants with more specific mechanisms of action. Compounds like beta-glucans, cytokines and different PAMPs (e.g. TLR ligands), alone or in combination, are now studied as possible adjuvant candidates in fish vaccines (Tafalla et al., 2013, Dalmo and Bøgwald, 2008, Thim et al., 2014, Thim et al., 2012, Carrington and Secombes, 2006, Fredriksen and Grip, 2012, Rivas-Aravena et al., 2015).

TLR ligands as potential vaccine adjuvants are covered in section 4.4.

3. Fish immune system

Fish is the largest class of vertebrates and can be divided into jawless fish and jawed fish, and the latter can be further divided in to cartilaginous fish (e.g. sharks) and bony fish (e.g.

teleosts) (Hitzfeld, 2005). Bony fish (Osteichthyes) possess most of the components in the immune system associated with the mammals. Teleosts are a branch of bony fish to which most of the economically important species belong (e.g. salmonid, carp, and tilapia species). Zebrafish (Danio rerio), a species important in research, also belongs to the teleosts. In contrast to higher vertebrates, fish are free-living organisms from early embryonic stages of life in which they depend on their innate immune system for survival (Uribe et al., 2011).

3.1. Fish immune organs and cells

The immune organs of teleost fish differ from other vertebrates. Thymus, head kidney, and spleen are the main lymphoid organs in teleost, while bone marrow and lymph nodes are lacking. The head kidney (anterior part of the kidney) performs important immune function and is considered equivalent to the bone marrow in mammals. It also functions as a secondary lymphoid organ along with the spleen (Kaattari and Irwin, 1985, Kibenge et al., 2012). Gut-associated lymphoid tissue (GALT) is well developed in teleosts (Salinas, 2015), and a unique interbranchial lymphoid tissue (ILT) has been identified in salmonids (Koppang et al., 2010).

Teleost fish possess most of the immune cells known from the mammalian immune system; neutrophils, monocytes/macrophages, eosinophils, non-specific cytotoxic cells (similar to mammalian natural killer cells), and T and B lymphocytes (Rauta et al., 2012,

(23)

23

Secombes, 1996, Whyte, 2007). Epithelial cells may also be involved in innate defence in fish (Whyte, 2007). Dendritic-like cells have also been found in some fish species and may together with macrophages and B cells act as antigen-presenting cells (Bassity and Clark, 2012, Lugo-Villarino et al., 2010, Rauta et al., 2012). The main phagocytic cells in in fish are neutrophils and macrophages (Secombes and Fletcher, 1992). Respiratory burst and nitric oxide have been demonstrated in fish phagocytes (similar to homologous responses induced in mammalian phagocytes) and have been shown to be critical effector mechanisms in limiting the growth of fish pathogens (Neumann et al., 2001).

3.2. Components of fish innate immune system

The innate immunity is a fundamental defence mechanism in fish due to limitations of the adaptive immune system (Whyte, 2007). The physical barriers consist of skin, scales, and gills and represent the first line of defence against pathogens. The mucus covering the skin contains various components (e.g. lysozymes, IgM, antibacterial peptides, complement proteins, and lectins) which inhibit entry of pathogens (Uribe et al., 2011).

Fish secrete a wide range of antimicrobial peptides (i.e. in the saliva, mucus, and circulatory system) that play major roles in the innate immune system and protect against a variety of pathogens (Rajanbabu and Chen, 2011). The complement system of fish seems to have all of the three complement activation pathways known from the mammalian system. Compared to other vertebrates, fish possess a number of genes encoding several complement components that are structurally and functionally diverse, indicating the importance of this system in a rapid response against invading pathogens (Plouffe et al., 2005, Zhu et al., 2013)

A number of TLRs have been identified in teleosts, and a more detailed description of mammalian and fish TLRs is presented in section 4. Other PPRs found in fish are RLR and NLR families. The three mammalian members of the RLR have been identified in several fish species (e.g. rainbow trout, Atlantic salmon, and zebrafish); RIG-1 (Biacchesi et al., 2009, Nie et al., 2015), melanoma differentiation-associated gene 5 (MDA5) (also known as IFIH) (Sun et al., 2009, Chen et al., 2015, Chang et al., 2011a), and laboratory of genetics and physiology 2 (LGP2) (Chang et al., 2011a, Chen et al., 2015). In mammals, RLRs are responsible for detection of cytoplasmic viral RNA and they appear to be involved in antiviral immune responses in fish as well (Kawai and Akira, 2008, Poynter et al., 2015). NLRs are also cytoplasmic receptors and sense bacterial cell wall components in

(24)

24

mammals (Kanneganti et al., 2007). They are also present in several fish species (e.g.

rainbow trout and zebrafish) and are most likely involved in antibacterial and antiviral defences (Laing et al., 2008, Chang et al., 2011b, Zhu et al., 2013). In zebrafish, the intracellular peptide glycan (PGN) recognition proteins (PGRPs) have been identified and may work together with TLR2 in recognition of PGN from bacteria (Chang and Nie, 2008).

Cytokines are a family of low molecular weight proteins that are secreted from immune cells (e.g. macrophages and lymphocytes) upon pathogen encounter to modulate inflammation and cope with pathogen infection. They can be divided into interferons (IFNs), interleukins (ILs), tumor necrosis factors (TNFs), colony stimulating factors, and chemokines (Savan and Sakai, 2006). In general, fish possess a repertoire of cytokines similar to mammals (Reyes-Cerpa et al., 2012) and the most characterised ones in fish are the pro-inflammatory cytokines IL-1β and TNF-α (Plouffe et al., 2005).

IL-1β is a pro-inflammatory cytokine, one of the first cytokines to be upregulated during an infection, and has been found in many fish species (Secombes et al., 2011). The fish IL- 1β share many of the characteristics with the mammalian counterpart, e.g. increases phagocytosis, chemotaxis, superoxide production, expression of important immune transcripts, leucocyte proliferation, and resistance to infection (Savan and Sakai, 2006, Plouffe et al., 2005, Hong et al., 2001, Peddie et al., 2001, Secombes et al., 2011). IL-1β activates target cells by binding to IL-1 receptors (IL-1R). Genes similar to the human IL- 1R gene have been identified in rainbow trout and Atlantic salmon and were upregulated during LPS treatment (Sangrador-Vegas et al., 2000, Subramaniam et al., 2002). TNF-α has also been identified in numerous fish species and differ from the mammalian counterpart in the presence of multiple isoforms in some species (e.g. zebrafish and rainbow trout) (Reyes-Cerpa et al., 2012). Expression of TNF-α has been shown to increase during LPS stimulation in several fish species (Plouffe et al., 2005). While recombinant trout TNF-α has been shown to induce increased migration and phagocytic activity in trout head kidney leukocytes (Zou et al., 2003), in several other fish species the in vitro effects of TNF-α were surprisingly weak (Reyes-Cerpa et al., 2012). IL-6 is another important pro-inflammatory cytokine in the early mammalian immune response against pathogens, but little is known about its functions in fish. IL-6 can be upregulated by mimics of infection and seems to have similar effects as IL-1β in rainbow trout and

(25)

25

zebrafish (Costa et al., 2011, Varela et al., 2012). However, IL-6 downregulated IL-1β and TNF-α in trout head kidney macrophages, suggesting a potential role in limiting host damage during inflammation (Costa et al., 2011). IL-10 is known to be anti-inflammatory in humans and the gene is associated with suppression of Th1 response (Brocker et al., 2010). IL-10 has been identified in many fish species and while its role is not clear, it has been associated with mechanisms of immune evasion by IPNV in Atlantic salmon (Reyes- Cerpa et al., 2014). Chemokines are chemotactic cytokines that are involved in recruiting immune cells to the infection site. The most studied fish chemokine is IL-8, which has shown chemo-attractant properties in rainbow trout (Omaima Harun et al., 2008) and has been tested as an adjuvant in a VHSV DNA vaccine (Jimenez et al., 2006).

In mammals, interferons (IFNs) are the first line defence against viral infections. The large number of IFNs identified in vertebrates are grouped in type I (e.g. IFN-α and IFN-β), II (IFN-γ), and III (IFN-λ). Type I and III induce specific antiviral immune defences, while type II is involved in promoting cell-mediated immunity (Zou and Secombes, 2011). The nomenclature of fish IFNs has been controversial and several names and classifications exist. The type I IFN family of fish contains at least the four subtypes IFNa, IFNb, IFNc, and IFNd (Zou and Secombes, 2011). Atlantic salmon possess 11 virally induced IFN genes in a multiple gene cluster: two IFNa (IFNa1 and IFNa3), four IFNb (IFNb1–b4), and five IFNc (IFNc1–c5) (Sun et al., 2009). In addition, IFNa2 and IFNd has been found in Atlantic salmon, but outside the multiple gene cluster (Svingerud et al., 2012). Zebrafish also has an IFN gene cluster encoding IFNa1 (also called IFNϕ1/IFN1), IFNc1 and IFNc2 (also called IFNϕ2/IFN2 and IFNϕ3/IFN3, respectively), but no IFNb. In addition, a zebrafish IFNd1 (also called IFNϕ4) has been identified (Zou and Secombes, 2011, Hamming et al., 2011). Fish also possess homologues of mammalian type IIs (IFNγ) and these might be involved in both antiviral and –bacterial responses (Zou and Secombes, 2011). IFN expression is modulated by a family of transcription factors called interferon- regulatory factors (IRFs) which has been shown to exist in all vertebrates (Zhu et al., 2013). Type I IFNs work through IFN receptors to activate the Jak-Stat signalling pathway, of which all components have been identified in fish (Zhang and Gui, 2012, Levraud et al., 2007, Sun et al., 2014). This signalling pathway leads to expression of IFN-stimulated genes (ISGs) that exert numerous antiviral effector functions (Schoggins and Rice, 2011).

Multiple ISGs have been identified in fish (e.g. ISG15 and Mx) that have shown to be virus-induced and exert antiviral activity in several fish species (Altmann et al., 2004,

(26)

26

Robertsen et al., 1997, Røkenes et al., 2007, Langevin et al., 2013, Jensen et al., 2002, Zhang and Gui, 2012). Furthermore, members of the ISG protein family IFIT (interferon- induced proteins with tetratricopeptide repeats) have shown antiviral function in zebrafish (Varela et al., 2014).

3.3. Brief overview of adaptive immunity in fish

The adaptive immunity can be divided into cell-mediated and humoral immunity. Fish seem to have lymphocyte subpopulations similar to the mammalian B and T cells and possess many important genes related to an adaptive immune response: MHC class I and II, T-cell receptor, CD4, CD8, and immunoglobulins. The presence of cytotoxic T cells (CD8+ cells) has been suggested (Nakanishi et al., 2011). Moreover, cytokines that in mammals are considered signature cytokines for Th1, Th2, and Th17 responses, have been identified in fish, and thus suggest the presence of Th1, Th2, and Th17 cells (Laing and Hansen, 2011). The main immunoglobulin in teleost is IgM, which has a heavy chain similar to the mammalian B cells. Fish IgM has a tetrameric structure (as opposed to the mammalian IgM pentameric structure) and is the primary antibody in fish serum (Solem and Stenvik, 2006). Additional immunoglobulin isotypes have also been identified in fish:

IgD and IgT/IgZ (called IgT in rainbow trout (Hansen et al., 2005) and IgZ in zebrafish (Danilova et al., 2005)). IgT may be important in the GALT of rainbow trout, thus perhaps representing a functional analogue to IgA (Zhang et al., 2010). The existence of B cells with phagocytic and bactericidal activity has been suggested (Sunyer, 2012). Isotype switch has not been described in fish and the antibody response is generally known for being slow, having low affinity, and being temperature dependent (Sunyer, 2012).

However, there are several examples showing that fish are able to induce specific and strong antibody responses after pathogen challenge or vaccination, and that antibody levels can be used as a correlate of protection (Munang’andu et al., 2013, Solem and Stenvik, 2006, Fjalestad et al., 1996, Steine et al., 2001, Thim et al., 2012).

4. Toll-like receptors (TLRs)

4.1. Mammalian TLRs and their ligands

In human vaccinology, the TLRs and their ligands have been extensively studied. TLRs are, as previously mentioned, important receptors that sense invading pathogens (O'Neill et al., 2013). TLRs appeared in the early stages of evolution and have been conserved in both

(27)

27

invertebrate and vertebrate lineages (Medzhitov and Janeway, 2000). The first evidence of the Toll NF-κB-like signalling pathway was discovered when the Toll protein was found to be required for fungal resistance in fruit flies (Drosophila melanogaster) (Lemaitre et al., 1996) and a short time later the human homolog of Toll was found (Medzhitov et al., 1997). Since then, TLRs have been described in a wide variety of vertebrate species. Six major TLR families (TLR1, TLR3, TLR4, TLR5, TLR7, and TLR11) have been identified, and TLRs within a family recognise a general class of PAMPs associated with the family (Roach et al., 2005). TLRs are transmembrane proteins containing an extracellular recognition domain composed of multiple leucine-rich region (LRR) motifs, a transmembrane region, and an intracellular Toll/interleukin-1 receptor (TIR) signalling domain (named TIR because the similarity of the IL-1R signalling domains) (Botos et al.).

Upon binding a ligand, the TLRs are relocated into the lipid raft fraction of the cell membrane (Sadikot, 2012) before two TLRs dimerise (either heterodimerisation or homodimerisation) for firm ligand binding (Jin and Lee, 2008). The close proximity of the TIR domains of paired TLRs allows recruiting of TIR domain-containing adaptor proteins.

These adaptors are Myeloid differentiation primary response gene (88) (MyD88), Mydd88- adaptor-like (MAL, also known as TIRAP), TIR-domain-containing adaptor protein inducing IFN-β (TRIF, also known as TICAM1), TRIF-related adaptor molecule (TRAM, also known as TICAM2), or sterile α-and armadillo-motif-containing protein (SARM) (O'Neill and Bowie, 2007). The engagement of the adaptor molecules stimulates downstream intracellular signalling pathways. These pathways involve interactions between IL-1R-associated kinases (IRAKs) and TNF-receptor-associated factors (TRAFs) and will eventually lead to activation of transcription factors (nuclear factor kappa B (NF- κB), interferon regulatory factors (IRFs), cyclic AMP-responsive element-binding protein (CREB), or activator protein 1 (AP1)) that control hundreds of different immune-relevant genes. The intracellular signalling cascade is complex and several other factors are involved; however, not all of them can be described in detail here. The adaptor molecules involved, the signalling pathway induced, and the cytokine expression profile following stimulation of a TLR, depend on the type of pathogen and the TLRs recognising the pathogen, as shown in Figure 2 (O'Neill et al., 2013). The transcription factors activated lead to upregulation of pro-inflammatory cytokines (e.g. IL-β, TNF-α, and IL-6) involved in inflammation and/or Type I IFNs (IFN-α, IFN-β) involved in antiviral immune response (see Figure 2).

(28)

28

Figure 2: Overview of mammalian TLR signalling pathways. Figure from (O'Neill et al., 2013).

The diversity of the LRR folding defines the TLR binding specificities and will orchestrate the appropriate innate and adaptive immune responses against the specific pathogen.

TLR1, TLR2, TLR4, TLR5, and TLR6 are localised on the cell surface (TLR4 can also be found in endosomes) and recognise cell wall PAMPs, while TLR3, TLR7, TLR8, and TLR9 are localised in membranes of intracellular compartments (like endosomes) and recognise nucleic acids. Bacterial PAMPs are mainly sensed by TLR1, TLR2, TLR4, TLR5, TLR6, and TLR9, while viral PAMPs are sensed by TLR3, TLR7, TLR8, and TLR9 (Jin and Lee, 2008, Werling et al., 2009). An overview of mammalian TLRs and their ligands is presented in Table 2.

(29)

29 Table 2: Mammalian TLRs and their ligands

TLR Species Localisation Ligand examples

TLR1/TLR2 Human

Mouse Plasma membrane Bacterial triacylated lipopeptides Mycobacterial products

Porins

Synthetic triacylated lipopeptides (e.g. Pam3CSK4)

TLR2/? Human

Mouse

Plasma membrane Bacterial lipoproteins

Staphylococcus peptidoglycan Viral proteins (from certain viruses)

TLR3 Human

Mouse Endosomal membrane Viral double stranded RNA PolyI:C

TLR4 (+ MD-2 and CD14)

Human

Mouse Plasma and

endosomal membrane Lipopolysaccharide (LPS)

Lipid A derivatives (e.g. monophosphoryl A (MPL)) Viral proteins (from certain viruses)

TLR5 Human

Mouse Plasma membrane Bacterial flagellin Recombinant flagellin TLR2/TLR6 Human

Mouse Plasma membrane Bacterial diacylated lipopeptides Lipoteichoic acid

Synthetic diacylated lipopeptides (e.g. FSL-1, Pam2CSK4)) Zymosan

TLR7 Human

Mouse

Endosomal membrane Viral and bacterial single stranded RNA Thiazoquinolines

Imidazoquinolines (e.g. imiquimod)

TLR8 Human

Mouse Endosomal membrane Viral and bacterial single stranded RNA Thiazoquinolines

Imidazoquinolines (e.g. resiquimod)

TLR9 Human

Mouse Endosomal membrane Viral and bacterial CpG DNA DNA:RNA hybrids

Class A, B and C CpG oligodeoxynucleotides (e.g.ODN2006) TLR10 Human Plasma membrane Unknown

TLR11 Mouse Endosomal membrane Profilin, flagellin TLR12 Mouse Endosomal membrane Profilin

TLR13 Mouse Endosomal membrane Bacterial ribosomal RNA, vesicular stomatitis virus References (De Nardo, 2015, Oliviera-Nascimento et al., 2012, Bowie and Haga, 2005, Shi et al., 2011)

Mammalian TLR2 requires heterodimerisation with TLR1 or TLR6 and is involved in recognition of bacterial and fungal cell wall components. Homodimerisation of TLR2 has been suggested, but not confirmed (Oliviera-Nascimento et al., 2012). Several synthetic compounds that mimic bacterial lipoproteins (e.g. Pam3CSK3 and FSL-1) are well established agonists for mammalian TLR1/2 and TLR2/6 (Okusawa et al., 2004, Aliprantis et al., 1999, Ozinsky et al., 2000). TLR4 recognises Gram-negative bacteria via the lipid A part of LPS (Poltorak et al., 1998). The recognition of LPS also requires the co-receptors MD-2 and CD14, and the LPS-binding protein (LBP) facilitates the transfer of LPS to CD14 (Park and Lee, 2013). TLR4 is unique in being able to signal trough both Myd88/MAL and TRIF/TRAM. It uses the adaptor TRAM to recruit TRIF and induce IRF3 activation and IFN-β expression, but can also use MAL to recruit Myd88, which

(30)

30

leads to activation of NF-κB and AP-1 that will induce pro-inflammatory cytokines (Thompson and Locarnini, 2007). In addition, TLR2 and TLR4 have been shown to interact with viral components (e.g. viral glycoproteins) and may participate in in viral detection (Bowie and Haga, 2005).

TLR3 is responsible for sensing viral double-stranded RNA (dsRNA) (Alexopoulou et al., 2001) and is the only TLR that works via the Myd88-independent pathway. The adaptor TRIF is instead essential for TLR3-mediated signalling. TRIF interacts with TRAF3 to activate a kinase cascade that leads to the activation of IRF3, which induces expression of IFN-β. TRIF can also interact with TRAF6 to induce a late phase NF-κB and pro- inflammatory cytokine response (Thompson and Locarnini, 2007, Jensen and Thomsen, 2012). Polyinosine–polycytidylic acid (poly I:C) is a synthetic analogue of dsRNA that binds TLR3 and is extensively used to mimic a viral infection (Fortier et al., 2004). The natural ligand for human TLR7 and TLR8 is single-stranded RNA (ssRNA) (Heil et al., 2004, Lund et al., 2004). In addition, a group of synthetic antiviral-RNA-like compounds (e.g. imiquimod) work by binding TLR7 and TLR8 (Hemmi et al., 2002, Lee et al., 2003a).

Human TLR9 recognises bacterial and viral DNA, which typically contain un-methylated CpG oligodeoxynucleotides (ODNs) in higher frequencies than human DNA (Hemmi et al., 2000). TLR7, TLR8, and TLR9 signal trough Myd88 and TRAF6 that lead to activation of both NF-κB to upregulate pro-inflammatory cytokines and IRF7 to upregulate IFN-α and IFN-β (Thompson and Locarnini, 2007, Jensen and Thomsen, 2012).

4.2. TLRs in fish

The advances in fish genomic research during the last decade have led to the discovery of TLR genes in many species of bony fish. Ever since the first teleost TLR was discovered in goldfish macrophages (Stafford et al., 2003), about 20 TLR types (TLR1, 2, 3, 4, 5M, 5S, 7, 8, 9, 13, 14, 18, 19, 20, 21, 22, 23, 24, 25, and 26) have been identified in more than a dozen of teleost species (Rauta et al., 2014). The fish TLRs and the factors involved in their signalling cascade have high structural similarity to the mammalian TLR system.

Figure 3 shows a phylogenetic tree comparing full-length TLR amino acids sequences from Atlantic salmon, rainbow trout, and zebrafish with the human and mice TLRs. The tree shows the homology of the different TLRs between the different species, and that the TLRs make up six TLR families. Most vertebrate genomes are actually found to have at least one gene representing each of the six major TLR families (Roach et al., 2005).

(31)

31

Although fish TLRs have high structural similarity with mammalian TLRs and possess many homologues of the mammalian TLRs, they also have distinct features and differences. Some mammalian TLRs have yet to be found in fish and several non- mammalian and fish-specific TLRs have been identified (Palti, 2011). In many fish species (e.g. Atlantic salmon and rainbow trout), a soluble form of TLR5 (TLR5S) has been identified in addition to the membrane-bound form (TLR5M). Two putative soluble forms of TLR20 have also been found in Atlantic salmon (Lee et al., 2014). Other non- mammalian TLR genes found in several fish species are TLR14 and TLR18 that branches with the TLR1 family, and TLR19-26 that belongs to the TLR11 family alongside mouse TLR11-13 (see Figure 3). Some of these TLRs are unique to fish (e.g. TLR22), while others have only been found in aquatic animals (e.g. TLR14) (Rauta et al., 2014).

Orthologues of TLR6 and TLR10 seem to be absent from fish genomes, but TLR14 and TLR18 have been proposed as possible substitutes (Zhang et al., 2014). Zebrafish is one of the few fish species in which TLR4 has been identified (Meijer et al., 2004, Jault et al., 2004). Fish TLR21 has shown similarity to chicken TLR21, which is considered a functional homologue to mammalian TLR9 (Brownlie et al., 2009). Genome- and gene duplication events have been contributors to the diversity of the TLRs in fish and a number of duplicate multi-copy TLRs have been identified (e.g. Atlantic salmon TLR8a1, TLR8a2, TLR8b1, and TLR8b2) (Palti, 2011).

Unc93B1 is a chaperone that appears to be important for the trafficking of endosomal TLRs (TLR3, TLR7, TLR8, TLR9, and TLR11-13) within the mammalian cell (Gay et al., 2014). The gene has been identified in Atlantic salmon and zebrafish, and is thought to have a role in fish TLR signalling (Yeh et al., 2013, Lee et al., 2015). Most of the downstream molecules involved in TLR signalling have also been identified in fish and the pathways seem to be conserved. However, information on the functional importance of many of these genes is lacking (Rebl et al., 2010, Kanwal et al., 2014). Myd88 has been identified in Atlantic salmon, rainbow trout, and zebrafish, and seems to function similarly to the mammalian counterpart (Skjæveland et al., 2009, van der Sar et al., 2006, Iliev et al., 2011, Rebl et al., 2009). Furthermore, MAL, TRIF, IRAK4, TRAF6, NF-κB, IRF3, and IRF7 have also been identified in fish (Phelan et al., 2005, Brietzke et al., 2014, Iliev et al., 2011, Meijer et al., 2004, Kanwal et al., 2014, Purcell et al., 2006); but TRAM has not been identified in any fish to date (Zhang et al., 2014).

(32)

32

Figure 3: Circular phylogenetic tree of full-length TLR amino acid sequences of Atlantic salmon (Salmo salar, Ss), Rainbow trout (Oncorhynchus mykiss, Om), Zebrafish (Danio rerio, Dr), human (Homo sapiens, Hs) and mouse (Mus musculus, Mm). Bootstrap values based on 100 replicates are indicated on each branch. Accession numbers of sequences used to build the three are as follows:

Danio rerio (TLR1: AAI63271, TLR2: NP_997977, TLR3: NP_001013287, TLR4ba:

NP_001124523, TLR4bb: NP_997978, TLR5a: XP_001919052, TLR5b: NP_001124067, TLR7:

XP_003199309, TLR8a: XP_001920594, TLR8b: XP_003199440, TLR9: NP_001124066, TLR18:

AAI63840, TLR19: AAQ91317, TLR20a: AAI63786, TLR20f: AAQ91319, TLR21: NP_001186264, TLR22: NP_001122147); Homo sapiens (TLR1: NP_003254, TLR2: NP_003255, TLR3:

NP_003256, TLR4: NP_612564, TLR5: NP_003259, TLR6: NP_006059, TLR7: NP_057646, TLR8: NP_619542, TLR9: NP_059138, TLR10: NP_001017388); Mus musculus (TLR1:

NP_109607, TLR2: NP_036035, TLR3: NP_569054, TLR4: NP_067272, TLR5: NP_058624, TLR6: NP_035734, TLR7: NP_573474, TLR8: NP_573475, TLR9: NP_112455, TLR11:

NP_991388, TLR12: NP_991392, TLR13: NP_991389); Oncorhynchus mykiss (TLR1:

NP_001159573, TLR2: CCK73195, TLR3: NP_001118050, TLR5S: NP_001118216, TLR5S:

P_001117680, TLR7: ACV41797, TLR8a1: ACV41799, TLR8a2: ACV41798, TLR9: ACC93939, TLR19: CCW43211, TLR22a1: NP_001117884, TLR22a2: CAI48084); Salmo salar (TLR1:

AEE38252, TLR3: AKE14222, TLR5M: AEE38253, TLR5S: AEE38254, TLR7: CCX35457, TLR8a1: NP_001155165, TLR8a2: CCX35458, TLR8b1: CCX35459, TLR8b2: CCX35460, TLR9:

ABV59002, TLR13: NP_001133860, TLR18: CDK60413, TLR19: CDH93609, TLR20a:

CDH93610, TLR20d: CDH93613, TLR21: CDH93614, TLR22a: CAJ80696, TLR22a2:

CAR62394).Figure made by M.Arnemo with CLC Main Workbench7.

(33)

33

4.3. Ligand specificity of fish TLRs

Many functional studies have failed to detect ligand specificities for fish TLRs. Direct ligand specificity has only been reported for a few TLRs, using in vitro reporter assays to detect ligand recognition (Ribeiro et al., 2010, Matsuo et al., 2008, Tsujita et al., 2004, Yeh et al., 2013). However, numerous published stimulation and gene expression analyses discuss possible ligand specificities and roles of fish TLRs during infection. The studies concerning Atlantic salmon, rainbow trout, and zebrafish will be reviewed here.

Since TLR6 has yet to be found in most fish genomes, TLR1 is the most likely partner for heterodimerisation with TLR2 in fish (Pietretti and Wiegertjes, 2013). Few reports are available on salmonid TLR1 and TLR2, but bacterial infection has shown to upregulate TLR1 in vitro (Salazar et al., 2015), while Pam2CSK4 and Pam3CSK4 (classical ligands for human TLR2/6 and TLR1/2, respectively) seemed to have a lower potential for inducing TLR and cytokine expression (Palti et al., 2010b, Purcell et al., 2006). Zebrafish TLR18 branches with the TLR1 family and was upregulated together with TLR1 and TLR2 during M.marinum infection (Meijer et al., 2004). Atlantic salmon, rainbow trout, and zebrafish TLR3 have been cloned and characterised by expression analysis, which demonstrated upregulation of TLR3 and type I IFNs following infection with aquatic viruses or poly I:C stimulation (Phelan et al., 2005, Vidal et al., 2015, Svingerud et al., 2012, Purcell et al., 2006, Rodriguez et al., 2005, Jensen et al., 2002, Dios et al., 2010).

This indicates conservation of TLR3-signalling pathways as well as involvement in antiviral immunity and binding of dsRNA.

The high tolerance of LPS in fish was for a long time explained by the lack of TLR4 in most fish genomes (e.g. Atlantic salmon and rainbow trout). This explanation was challenged when two TLR4 genes (TLR4ba and TLR4bb) were identified in zebrafish (Meijer et al., 2004, Jault et al., 2004). However, the apparent absence of CD14, MD-2, and LBP from all fish genomes may play a role (Pietretti and Wiegertjes, 2013). Although zebrafish tolerate high doses of LPS, LPS has been shown to exert multiple biological effects (Novoa et al., 2009, Swain et al., 2008). It seems that zebrafish are responsive to LPS through a TLR4-independent pathway, thus suggesting an alternative LPS induction pathway in fish (Sullivan et al., 2009, Sepulcre et al., 2009). Aedo et al. (2015) proposed that TLR5M and TLR9 may play a role in detecting LPS in rainbow trout (Aedo et al., 2015). Conservation of flagellin binding by TLR5 has been suggested in rainbow trout,